Podcast 012 AI in Marketing
Episode 012 AI in Marketing Transcript
Heather M.: Welcome to The Modern Polymath where we discuss topics in technology, economics, marketing, organizational behavior, market research, human resources, psychology, algorithms, higher education.
Heather M.: Hey, podcast universe. Thanks for tuning in. On today’s episode of The Modern Polymath, we’re going to talk about how AI is being used within marketing and ways in which marketers are finding it useful to help with things like efficiency, marketing spend, customer service, and a whole lot more. As we are nearing the holidays, this will be our last episode for 2019. We’ve thoroughly enjoyed making this podcast and have been amazed at how many of you are actually listening. We thank you very much and wish you and yours a happy holiday season. Now, let’s get this podcast started.
Will C.: Okay, so AI in marketing. It’s a cool new topic. Marketers are marketing their AI and technologies, but AI really just looks like it’s a new feature of the marketing technology stack. What technologies would you say make up the current marketing stack?
Heather M.: Yeah, so content management systems, any advertising technology that places ads for you, email platforms, which also could be known as marketing automation platforms, any tool that’s giving you insight or any kind of analysis. Also, any kind of experience optimization, so if you’re doing A/B testing or something like that. Social media is a big one. Also, customer relationship management systems. That’s usually one that a marketing team may use in conjunction with the sales team so that everybody is communicating on the clients at real time. Then also SEO, search engine optimization. That’s a huge part of how AI is working behind the scenes.
Will C.: Yeah. With that as the new stack, we would probably even think of a lot of those different technologies, each one having their own layer of AI.
Heather M.: Absolutely. I used to use Marketo at one of my jobs and a cool thing that it did was whenever you set up an A/B test and you finally determine the winner, it would just go ahead and automatically start sending that winning email to everyone so you didn’t have to manually do it.
John-David M.: Placing ads on Facebook. People know that’s important or whatever it may be, but there’s a lot going on behind the scenes that builds this into it. I think that’s one of the important things about this topic is to understand that it’s not understanding the technical elements of it as much as understanding that it is running in the background and most of what’s there now and will continue to be more and more prevalent. By understanding how that fits together, you can better align your marketing stack and put together a strategy that makes sense while also in some ways being able to mitigate how much extra work you have because you’re able to automate things.
Heather M.: Staying on the marketing automation train for just another minute, one great thing that people can do using marketing automation and the AI that’s built in behind it is campaign optimization. If you have a contact or a lead in your system and you want to make sure that they get the right message at the right time based on what stage they are in in their journey with you, you would then set up certain triggers that would fire off based on behaviors that they took.
Will C.: Certain behaviors that they took within interacting with your website or a certain sales agent or at a certain time period as to keep your company fresh in their mind or keep that, “Hey, you still have that in your cart. You sure you don’t want to buy it? We’ve got a sale coming up on 10/10/10. Please buy this.”
Heather M.: Right. Exactly. That’s one great way that you can really utilize the ease that AI provides to you in terms of not having to manually sit and refresh a list to see if something has happened. Now AI allows us to automatically track and trigger certain other events to happen after something does take effect. Also, personalization. We all have gotten emails where it has your first name either in the subject line or in the email or sometimes people use really awesome graphics where their videos and animated and are able to put your name in there and have them show up. Some people get really creative.
Heather M.: But even recommendations that are made, all of that is tracking basically your preferences, your likes, interests based on the data that a company has been able to collect on you while you’ve been interacting with them so that they can recommend to you things that they think you would like. Not random things.
Will C.: Yeah, it’s just behavior analytics on your website, whatever they’re going to do, their click behavior, what they browse for. I mean, I don’t mind getting personal ads. I don’t mind getting real products pushed towards me that I would probably like even if they’re new and I never heard of this company. I mean, you have Google. You can Google around and see if the company’s legit or not, but I mean, that type of personalization like I have no problem with.
John-David M.: Yeah, it’s like you’re going to get ads. You might as well be interested.
Will C.: Yeah, might as well be accurate. Yeah.
John-David M.: I mean, how many things that we bought off Instagram ads and things like that-
Will C.: Yeah, for sure.
John-David M.: -were correctly targeted towards us?
Heather M.: Mm-hmm (affirmative). I actually bought something this morning off of an Instagram ad.
John-David M.: Well, that’s new.
Heather M.: Another one that, again, marketers should understand is using artificial intelligence but customers may not understand that this is using artificial intelligence are the chat bots that are on several websites. Anytime you go and you have the little box pop up with either a person’s face or something that says, “Hi, can I help you,” that’s not actually a person, for all of those of you who didn’t know. That’s not really a person. That is a chat bot. I will say there is human intervention in there at some point, but at that first interaction, that marketing team has basically designed out whatever script it needs to that would match whatever question you would type in there. Say you’re looking for a…
Heather M.: You go to Home Depot’s website and you’re looking for a certain color of paint and something pops up or a chat bot pops up. They ask, how can I help you? You say I’m looking for a Benjamin Moore Paint or whatever it is that they carry. Then at least now the marketing team knows what to tell you about what you need because paint has popped up as a keyword. They can start sending you all the information regarding that. Just at the point where the bot can’t answer you anymore, that’s where then there’s usually human intervention that steps in.
John-David M.: Yeah, because the AI is in there logging the conversations that have happened in the past and the answers that you get and continually learning to be able to give better answers.
Heather M.: Yeah, and to be more efficient as well because obviously everything is about the customer experience now and receiving good customer service. If chatbots can cut down the amount of time that a consumer has to spend with your company, especially dealing with a certain issue, whether it be negative or even making a direct purchase, if you can make it super quick and easy for somebody to buy from you rather than having to call someone to place that order, they’re going to be a lot more likely to buy.
Will C.: Yeah, and there are efficiencies on the back end for the company, right? They don’t have to staff that person to take that call in whatever call center they’re going to purchase services from.
Heather M.: Absolutely.
John-David M.: Not to mention you’re also logging data every time you go through that process. Whereas if you’re relying on an individual to log that, you may not get it. You’re able to get smarter and better. I mean, it wasn’t too long ago that it was so obvious that every time you got a call from a chatbot or a bot, you knew it was. Now we’re getting fooled by them a lot. That’s not necessarily a good thing, but it does illustrate the ability to be much more intelligent and efficient.
Heather M.: Yeah, and that’s a great point and something that I was about to just lead into is that chatbots will get better as the artificial intelligence piece of that in terms of speech recognition and natural language processing. As that continues to develop and get better, then you’re going to be served with better information because it’s learning. But you’re right, you can’t really rely on humans to track that. I mean, how many times have we asked for someone on the sales end of a call to ask them how they heard about you or something. That question always gets skipped.
John-David M.: I’d rather not talk about that.
Heather M.: We’ve talked about ads and being served ads, but from a marketing standpoint, a marketer definitely cares about ROI on that ad spend because it has to be justified, but one of the best ways that you can actually measure that is through attribution. Sometimes that is easier said than done, being able to tell exactly what ad, what email, what trade show or whatever it may be that was the first touch, last touch, third touch that possibly got that person to buy. Machine learning now, which is part of AI, can help identify the customer touch points to remove that human bias. We may see some pattern that we want to see, whereas the data can actually tell us what truly is.
John-David M.: Different phases you have to attribute where they are in that to better qualify them and then know where to spend your time and where not too. That’s one of the most important things is everybody that you… From a marketing perspective, everybody you go after cost money and there’s a cost to every interaction with that individual or business. The more efficient you can be through proper attribution, the less time and resources you’re going to waste on people that aren’t a good fit and the more you can really focus on those that are and give them more of a personalized experience.
Heather M.: One of the companies out there that’s actually doing something with this right now is LinkedIn. They actually acquired Drawbridge for better ad target and attribution within the LinkedIn platform. This is going to make it a lot more efficient for advertisers whenever they’re actually looking to target these customers because there’ll be able to leverage Drawbridge’s specialties as well as LinkedIn. Another cool place where we’re really starting to see the use of natural language processing is in call tracking and analytics. There’s a company called Dialpad that just introduced a measurement for customer sentiment, so whether it was positive, negative, neutral, whatever other categories they may be looking to measure.
Heather M.: This is really interesting because I think back to when I was working with an inside sales team and it would have been wonderful to have a record of the calls that they were having with the clients so that you could try to figure out if it was a positive call, what were they talking about? If it was a negative call, what were they talking about so can highlight the issue. Was it pricing? Was it they got the wrong order? Was it that it was a late delivery? Try to narrow down on what the issues are, good or bad. Something like this, especially if you’re not getting good online reviews, this would be a great thing to use to help improve your customer service.
Will C.: Yeah, and being able to join those in a call type situation as well as parsing through your Amazon reviews, doing a sentiment analysis quickly and kind of getting a view of the landscape to look closer at what issues are actually popping up, what do you really need to pay attention to.
Heather M.: Yeah, and on that front talking about what you need to pay attention to, I know that social media for some folks in the marketing world is something that some people take it seriously and some people don’t. I think that all marketers recognize the importance of it, whether they think it’s a great platform or a bad platform, but they do recognize the fact that there is a large audience there, big reach. But one of the things that you have to do to be able to make it a continued successful platform is to pay attention to those algorithm changes and learning to adjust your marketing efforts to make sure that you continue getting exposure to your customers on their feeds and that the platforms just start picking you up even for keywords, trending items, that sort thing.
Will C.: Yeah. On that front, I mean, you need to pay attention to the algorithms changing because that’s your marketing spend right there, right? If the algorithm changes and you start marketing people at 2:00 AM where most of your customers are asleep, it’s like, oh great, we’re wasting money because I didn’t understand that the algorithm changed last week for whatever reason.
Heather M.: Right. I mean, there’s a lot of fronts where marketers are interfacing with AI on a daily basis whether or not they recognize it to be under the umbrella of AI. Of course, we just listed out a lot of different examples and different uses. But one of the big things that people should or that marketers should also remember is that a lot of consumers aren’t necessarily quick to adopt the use of AI. Some of them would prefer to work with a human and it may also depend on the type of business that you have, the service that you’re offering, whatever that may be.
Heather M.: One thing to remember is that if you are going to use AI, you just need to communicate transparently and make sure that you’re building trust with your customers because the worst thing that can happen is they lose trust over a robot and not an actual salesperson.
John-David M.: I mean, again, like it can be used for very good, like we’re talking about with the Instagram ads. There’s a lot of different ways things can go, but it can very much enhance the customer experience, the offerings that people are able to give you, and the ability to get feedback from you to improve what they’re offering. There’s a lot that you can learn. We say this with a lot of things, but with anything powerful, there’s a good and there’s a bad, there’s an upside, there’s a downside. AI is powerful. Data is powerful. You can use it for good. You can use it for bad.
John-David M.: But as a company with the realization of the importance now of the customer experience and the ability for average customer B2B or B2C to be able to get information easily through Google and find whatever they need, you’ve got to do right by your customers. AI is a very powerful way to do that, to use it in a way that can create a more loyal customer base that will guarantee you distribution when you roll out a new product because you’ve gotten feedback from them and you have your clients have buy in into the process of developing the products and how you articulate that. All these different things are possibilities if used correctly. However, if you violate the trust of your customer base using these types of things, it’s probably going to come back to bite you.
Heather M.: Definitely.
John-David M.: As people get more savvy, they understand more how this works, maybe not on the technical side, just intuition starts to kick in as they see it enough and enough and enough. They’re going to become wise to that.
Heather M.: Yeah, and that’s where that whole creepy factor comes into play where it’s a really fine line to tow kind of because you want to deliver the best and most personalized experience to your customer, but you don’t want them to feel like you have a microphone in their room all the time or that you really are listening through their phones. One thing to remember is that no matter how fancy of a technology or AI marketing stack you have, so all the different programs and platforms that you’ve implemented and are using, no matter how great they are, if you have garbage data in, it’s putting garbage data out.
John-David M.: Absolutely.
Heather M.: This is probably one of the most painful points for actually implementing AI in any company is making sure that you’ve got the right data flowing in, that it’s not siloed, that everyone is able to see the same data, and of course, that it’s not being changed by inappropriate levels of folks for reasons when it shouldn’t be. What we like to say is that really you need to understand what type of AI your company has now to support the data that you have now, enhance that data source and the processes where you need to, and then implement the actual technology piece to it.
John-David M.: Yeah. I mean, it goes back to some of the things that we’ve talked about in prior podcasts and have been written up in multiple journals and stuff about is having a data culture and understanding the importance of having the right data because if you don’t… Data’s extremely powerful like we’ve said before, but it can come back to bite you if it’s not good data and it’s not used correctly. There’s an immense power that AI can drive in terms of the ability to effectively market to people. But if the data is not good, if your models aren’t good and you’re relying on… It’s just like you get what you pay for in a lot of cases and people sometimes try to build models on the cheap. That can cost you a lot more money than it saves you.
John-David M.: If you’re basing your assumptions on a model that is flawed or you’re basing assumptions on bad data and you’re allocating spend and other resources to the “insights” that you got out of that, that’s extremely dangerous if it’s leading you astray. You have to do it correctly. I can’t tell you the number of times we’ve come in and fixed somebody’s model or looked at their data and seeing just how messy it was and they’re trying to derive strategic insights out of something that is telling them nothing but noise really. That’s very dangerous. You have to do it correctly and that’s true of marketing and management, finances, anything else.
Heather M.: Absolutely. Part of having a data culture means that your marketing team is data fluent as well because a lot of marketing resources are spent as part of a company’s overall budget should be anyway because it’s a very important part of a business. But you can’t just go justify spending a million dollars in ad spend on either certain channels when you have no data telling you that you should be and especially if the data that you’re getting in is telling you that those channels aren’t working. You need to have their systems set up so that you know if it’s working or not. That goes back to the marketing attribution. What is actually working and what is not so that you can continue to optimize. The beauty of it is is that you have all these AI tools to do it with.
Heather M.: They’re making everything so much more automated or at least to where you can build something and let it do it again for you. Maybe you have to touch it occasionally, but for the most part you can usually set something up, tell it what to measure, and then just pull your reports. That is taking people beyond where I think they ever thought that they could be because now we have all this at our disposal and at our fingertips and seeing the cool things that people are doing with it, like dynamic product pricing or sales forecasting or even the new area of voice search that people are going to start having to figure out how to make their websites friendly for voice search because that’s becoming a more and more popular way of consumers to find things.
John-David M.: But one of the things that’s missed in the midst of all that, and I think an important thing to point out, and we’ve said this many times, and I’m sure we’ll say it many times in the future, is you can’t just throw tools at a problem and hope that it solves it. You have a limited amount of resources both in terms of a budget for tools as well as the people that can actually implement them. One thing that you have to make sure as you add AI in, as you incorporate it into your business no matter where it is, it needs to be properly aligned with the other technologies that you’re using. The data needs to be good and clean. You’ve got to set it up right. Adding another layer of complexity onto something that’s already…
John-David M.: If you’ve tried to make a stew and you throw a bunch of random ingredients in there, it’s very unlikely that another random ingredient is going to make it better. Sometimes it’s just the KISS principle, keep it simple. You really want to make sure that as you layer in a new layer of complexity, anything data-driven, especially when it comes to something like AI, which is continually learning off the data that you have and are collecting and the interactions that are taking place on whatever the platform may be, it needs to be integrated into a larger strategy.
John-David M.: I know one of the next points we’re going to talk about and I think is important to hit on is in setting this up, marketing teams as well as other teams need to look at getting true data scientists in there to design this from really an architectural perspective. How do the pieces fit together? How does the data all… We’ve seen so many times where you have data being collected in multiple different systems and the systems don’t talk to each other. You have one department has really good information on this, marketing has great information on this, sales on this, finance on this, ops on this. It’s just like, oh my god, if we combine all that together, we’d know a lot. That’s one thing that you can hire internal data scientists.
John-David M.: Because of the market and the way things are, at least at the beginning stages, hiring someone to come in and help you from a consulting standpoint to get this set up correctly. Because also the amount of talent out there in the data science field is very, very much under serving the market needs as a whole, therefore, data scientists are very much in high demand. You’ve got to make sure you do it right on the good. If you get it set up, if the foundation is right, it’s good. If you build a house on a bad foundation, I think it might come crumbling down. If you want to add another story on top of it, that’s not going to make it anymore secure. If you add in more height over the Tower of PISA, it would probably topple on over, right? The foundation wasn’t good to begin with.
John-David M.: You got to make sure that part is done correctly. AI should be a secondary priority after making sure your data is all aligned, your systems are talking to each other.
Heather M.: Agreed. Just because you bought a subscription to Power BI or Tableau doesn’t mean anything if you don’t have somebody on your team that can put the data in and work it. You need someone, a data scientist, who is pulling all that great sales data, all that great marketing data, all that great finance data, product data, whatever it may be, together into Tableau or Power BI so that you can actually see something out of it and make something a more digestible narrative that helps to drive the strategy, not just to give you random updates of data points that don’t really mean anything.
Will C.: They’re arbitrary, which is something we talk about, which is defining what you’re trying to measure. You need to know the correct KPIs that you’re going to play with before you can even start analyzing the data, right?
Heather M.: Absolutely.
Will C.: That’s probably an iterative process of updating which KPIs are giving you the best lens to view your ROI as a whole for the marketing. But I mean, defining what you’re trying to measure what data sources you have, then you can start playing with the tools and see like optimize your workflow.
John-David M.: We’re going to talk about fallacies coming up soon in terms of critical thinking. That’s going to be one of our episodes, but this happens a lot. One specific fallacy happens a lot in businesses is the Texas sharpshooter fallacy. The fallacy is the guy shoots on the side of a barn and he shoots his rounds off and then he goes and he draws a circle around his target. You can’t hit a target you don’t define. Now, if you don’t care what it is, if you’re trying to shoot the side of a barn and you had something five feet to the right of where you actually hit several times, but you can circle the thing and say, “I hit it.” Great. If you can fool the audience that way, that’s a fallacy.
John-David M.: But if you try to pull that fallacy in with your business, it boils down to dollars and cents and that’s going to catch up to you. You have to have a clearly defined target. You have to make your data and your systems work for you. AI is a powerful, powerful tool. As Heather said, there’s so many user-friendly types of AI out there that can make you so much smarter and more efficient as long as you set it up correctly.
Heather M.: Absolutely. I think that’s really the key message is that AI can be super effective, make your teams so much more efficient on many different levels, but making sure that you have it set up correctly from the get-go. But one of the ways that I like to think about it is that everything that the marketing department does, whether their KPI… Obviously you’ll work to work off of your marketing KPIs, but your marketing KPIs should be set off of your company KPIs and making sure that what you’re trying to do aligns with the company’s strategy.
Heather M.: If you’re able to then implement AI into these certain areas to help you either spend more efficiently or gain more market share in a particular area, you can use AI, machine learning, all these things to either cull through data you already have, help you collect data you don’t have, help you reach customers you’ve never reached before, all the while taking a lot of that work off of your plate.
John-David M.: Yeah, it’s possible. Just be very intentional about how you set it up and you really can do a lot with that.
Will C.: Yeah. I mean, AI, you can start targeting… Like you were saying, you can reach new customers. You implement some simple AI rule-based stuff. AI can start curating a message to where it reaches those new audiences and resonates with them to where they become a loyal user. It’s an automated task because you’re just letting AI run with it, right? Those types of things, I mean, yeah.
John-David M.: Yeah.
Heather M.: Exactly.
John-David M.: You can keep serving them automated messages that help pull them to the funnel, keep them engaged with your product. We all see it. Okay, I have something in my cart and I didn’t close on that, right? I just literally got an email from J. Crew on a Christmas gift I’m buying and it says low inventory on that. Now, that’s pretty simplistic AI. But nonetheless, it’s reminding me that I was interested in this item and it’s telling me that if I don’t order it soon, I’m not going to get it. Well, if I really want that, that’s good information to have.
Will C.: Yeah. With a little economic nod. Wink, wink.
John-David M.: We actually have 7,000 of these in the back.
Will C.: Is it really almost gone or did we just need you to buy this?
Heather M.: To your point, John-David about we’ve all seen this, I think that’s what’s interesting about AI in marketing is that it’s not new truly. Email marketing and that type of marketing automation and even Google Ads and Facebook Ads, social ads, whatever it is, it’s been around for at least a decade now. It’s not new.
John-David M.: Amazon recommendations and plugging your cell phone with the keywords and make sure that you can get found by the product you’re trying to sell off of.
Heather M.: All of that is marketing.
Will C.: It is. All of it.
Heather M.: It is straight up user experience marketing. I think something that people should keep in mind is that yes, the tried and true AI that has been in use and in place for a while will probably still continue to be relevant, but you really need to stay on top of what’s coming out and how these things are changing and how the field of AI itself is developing. Because if you know that they are making strides in natural language processing, then you may want to invest a little more the next fiscal year on your voice search strategy, or you may want to invest more in a certain chatbot or call analytics.
Will C.: Yeah, that’s a great point, Heather, because the companies that are pushing AI the most right now and are making the most strides just happened to be the biggest ad agencies in the world, right? It’s Google. It’s Facebook, Instagram. Those are the people who are pushing AI, so you need to know how their algorithms are changing like you were saying earlier, but you also need to know how they’re using AI, right? How are they using it to target certain customers, put you in certain buckets as an advertiser to then get pushed? I mean, it gets pretty complex because then it’s like AI on your end, AI on their end. But I mean, it can be great at like you’re saying.
John-David M.: But I want to say on the flip side of that, you can look at a Google or a Facebook and think that you can take your 50 person, two and a half mil a year company and get yourself to that level of sophistication. But in anything, you’ve got to prioritize it and break it down because I have seen personally a lot of companies, okay, AI is important. Data’s important. Great. Let’s throw a bunch of money at it and hope it works out. All right, so you want to get in shape. You want to get a six pack for the beach. You don’t just go from being a hundred pounds overweight with a terrible diet, eating Cheetos and downing Mountain Dew and never doing anything physical to having a six pack. You got to join a gym. You got to do cut back on the bad food.
John-David M.: Eat more good food. You have to make a dedicated intentional effort towards doing something, to build a foundation. You download the apps that’ll actually work for you and use them and get in there and do it and continually and iteratively make improvements in your lifestyle while you refocus your target. Getting closer and closer to being that level of efficiency or success that you predefined. But you can’t just set that end goal as being we’re going to be the most sophisticated AI company in our space without first building the building blocks and at each point iterating on those and becoming better and better. If you try to do it all at once, you’re not going to get there.
John-David M.: Now, I think that’s one thing to also remember from a consumer standpoint, if you’re say in a B2B space and you’re going to buy an AI solution, remember that AI companies are marketing themselves too and they’re going to tell you that you need to buy this because this is great and it’s whatever. That’s why I keep coming back to that point of the building blocks because yeah, it may be a great solution, but you can’t put the cart before the horse. They’re also marketers.
Heather M.: Yeah, exactly. Well, that wraps up this episode in our 2019 season. We hope that we’ve earned your favor to listen to us next season in 2020, which will cover some very interesting topics ranging from critical thinking and observable phenomenons to deep faking and math in society. Make sure to subscribe so you automatically get the next episode or stay in touch with us on Facebook, LinkedIn, Instagram or Twitter for updates for the next release date. We hope that you all have a wonderful and safe holiday season. Catch you later.
Awesome post! Keep up the great work! 🙂